EUVIMEDEUVIMED entdecken
Crossref

EEG-Based Cross-Operator Mental Fatigue Assessment for Human–Robot Collaboration

Yanghui Wang, Yong Zhou, Weidong Li, Duc Truong Pham

IEEE Sensors Journal · 2026-11 · 26 (21) · 33481-33492

DOI: 10.1109/jsen.2026.3736067 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Sensors Journal, 2026-11. Verfasst von Yanghui Wang und 3 weiteren Autor:innen. Die Arbeit nennt 36 Literaturangaben. Gefördert von: Anhui Provincial Key Program for Tackling Critical Challenges in Scientific and Technological Innovation, National Natural Science Foundation of China, Science and Technology Commission of Shanghai Municipality.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Mental fatigue can reduce attention, response capability, and operational reliability during human-robot collaboration (HRC). Electroencephalography (EEG) provides a noninvasive wearable method for monitoring fatigue-related brain activity. However, EEG-based fatigue assessment is challenging because multichannel signals are noisy, nonstationary, and variable across operators. To address these issues, this study proposes an EEG sensor-based temporal-frequency-spatial fusion framework for mental fatigue assessment in HRC. The framework integrates wearable EEG acquisition, signal pre-processing, feature fusion, and cross-operator adaptation. (i) A separable temporal convolutional network-graph convolutional network (STCN-GCN) is developed to extract temporal, frequency, and spatial features from EEG signals. The STCN captures temporal dependencies with reduced model parameters, while the GCN uses power spectral density (PSD) and Pearson correlation coefficients (PCC) to represent spectral information and channel connectivity. (ii) A domain adaptation strategy combining maximum mean discrepancy (MMD) with a domain obfuscator is designed to reduce inter-operator EEG variability and improve cross-operator fatigue assessment. Experiments show that the proposed framework achieves an average within-operator accuracy of 90.12% and an average cross-operator accuracy of 77.83%. An end-of-life product disassembly case study further demonstrates how EEG-derived fatigue indicators can support robot-speed adjustment and operator-rest decisions in HRC.

Förderung

  • Anhui Provincial Key Program for Tackling Critical Challenges in Scientific and Technological Innovation
  • National Natural Science Foundation of China
  • Science and Technology Commission of Shanghai Municipality
  • China Scholarships Council
  • Fundamental Research Funds for the Central Universities of China

Zitieren

APA

Wang, Y., Zhou, Y., Li, W., & Pham, D. T. (2026). EEG-Based Cross-Operator Mental Fatigue Assessment for Human–Robot Collaboration. IEEE Sensors Journal, 26(21), 33481-33492. https://doi.org/10.1109/jsen.2026.3736067

Vancouver

Wang Y, Zhou Y, Li W, Pham DT. EEG-Based Cross-Operator Mental Fatigue Assessment for Human–Robot Collaboration. IEEE Sensors Journal. 2026;26(21):33481-33492. doi:10.1109/jsen.2026.3736067

BibTeX
@article{Wang2026,
  author = {Wang, Yanghui and Zhou, Yong and Li, Weidong and Pham, Duc Truong},
  title = {EEG-Based Cross-Operator Mental Fatigue Assessment for Human–Robot Collaboration},
  journal = {IEEE Sensors Journal},
  year = {2026},
  volume = {26},
  number = {21},
  pages = {33481--33492},
  doi = {10.1109/jsen.2026.3736067}
}
Kostenlos starten →